From mboxrd@z Thu Jan 1 00:00:00 1970 From: Radim =?utf-8?B?S3LEjW3DocWZ?= Subject: Re: [PATCH 6/6] KVM: VMX: Add PML support in VMX Date: Tue, 3 Feb 2015 17:02:57 +0100 Message-ID: <20150203160256.GF19731@potion.redhat.com> References: <1422413668-3509-1-git-send-email-kai.huang@linux.intel.com> <1422413668-3509-7-git-send-email-kai.huang@linux.intel.com> <20150203151833.GD19731@potion.redhat.com> <54D0EBC8.6030301@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Kai Huang , gleb@kernel.org, linux@arm.linux.org.uk, kvm@vger.kernel.org To: Paolo Bonzini Return-path: Received: from mx1.redhat.com ([209.132.183.28]:46430 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932393AbbBCQDN (ORCPT ); Tue, 3 Feb 2015 11:03:13 -0500 Content-Disposition: inline In-Reply-To: <54D0EBC8.6030301@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: 2015-02-03 16:39+0100, Paolo Bonzini: >=20 >=20 > On 03/02/2015 16:18, Radim Kr=C4=8Dm=C3=A1=C5=99 wrote: > > (I see the same code in handle_ept_violation(), but wasn't that nee= ded > > just because of a hardware error?) >=20 > That was how I read it initially, but actually that means: "this > statement could be broken if the processor has that erratum". Thanks, that was a nice ruse for the original bug :) > >> +static void vmx_slot_enable_log_dirty(struct kvm *kvm, > >> + struct kvm_memory_slot *slot) > >> +{ > >> + kvm_mmu_slot_leaf_clear_dirty(kvm, slot); > >=20 > > (New slot contains dirty pages?) >=20 > New slots contain clean pages as far as the KVM dirty log is concerne= d. >=20 > In the case of PML, note that D=3D1 does not mean the page is dirty. = It > only means that writes will not be logged by PML. The page may thus > also have logging disabled. Yeah, it would be a problem if we had dirty pages at the beginning, but I don't think it is possible as was too lazy to check. (It's not important and I wanted to do this review today :)